AMELIE
AI-based decision support software for clinicians in diagnosing monogenic diseases

Genetic disease diagnosis is a challenging investigation, which in most cases is not concluded in one diagnosis attempt. Most cases require an ongoing continuous analysis of the literature which is being constantly updated by the scientific community, and may surprise with a resolution to an unresolved case. AMELIE enables the first analysis, as well as an automatic continuous analysis of a patient's case.

To start the diagnose, a clinician inputs the patient's genomes and his/her medical record. AMELIE processes the above information, using a machine learning approach and analyzes an extensive database of scientific literature to find the relelvant literature that explains the disease genetic causality.

AMELIE analyzes medical notes written as free text by doctors, and extracts automatically from the notes the patient's symptoms ranked by importance. The clinician is presented with the original text and the NLP analysis, with additional metadata that assists the clinician to select the relevant symptoms.

After AMELIE's processing, the clinician is presented with a list of ranked causal genes of the patient's disease, and associated literature per gene. AMELIE highlights per paper, the clinical symptoms which are an exact match, a missing or a approximate match to the patient's ones. These are amongst the tools to assist clinicians in selecting the most relevant papers that could assist in resolving the case.

First version of AMELIE case management showcases high level details and status per case, as well as allowing access to the full case and export to file of its details. The case management screen allows the configuration of continuous analysis per case.